During the recent Master’s golf tournament, I was able to make a connection between golf scores and z-scores. This relates to the connection between instantaneous z-scores from the live DLL, and the z-scores obtained from a 1-minute or longer assessment. It is typically seen that, even if a component is 3 or more sigma abnormal in the QEEG, when viewed using live z-scores, the values are 1 to 1.5 sigma lower. This has caused confusion and concern that the measurements are not correct. In fact, it is essential that an instantaneous z-score be lower than one from a longer epoch. The following analogy with a golf score helps to illustrate this fact.
The instantaneous z-score is like the score on a particular hole. In tournament golf, the score for any one hole is never very far from 2, 3, or 4, or 5. A score of 1 is possible (“hole in one”), and scores much above 6 are unusual (“double bogey”). In this case, you may be one or two points up or down (“birdie,” “bogey,” etc). However, the score on a single hole is never very far from the norm. You can think of a score of 3 as typically (“normal”) while each point above or below reflects something like one “sigma” up or down. Thus, in any hole, the “z-score” would almost never be much more than plus or minus 2. Even a really great golfer, cannot do that much better (or worse) on any individual hole. In other words, you cannot tell from the performance on one hole how good the golfer really is. In the Masters, there were several eagles, and even a hole in one. But the winner was the one who had the most birdies on average. That is, the one who sustained this “good but not great” performance for the most of the holes was the winner.
When we look at a QEEG postprocessed z-score, we are looking at the score for the entire round. Here we see the tally of each and every hole, and over 72 holes, the best performers are separated by only a handful of strokes. It is how well the player was able to achieve that “birdie” state consistently that separates the winners from the losers. So for a component to show up at 3 sigma, it is sufficient for it to have an instantaneous value of 1.5 or 2 sigma typically, but to sustain this for more time. In other words, by sustaining an instantaneous score on a moderately high side for a sustained period of time, the resultant z-score for the session can be very high. This is exactly the same as saying that, the more birdies you get, the better your final score.
So in the same way that a golf score consists of the combination of the scores for many individual holes, the total z-score for an evaluation reflects the combination of many instantaneous z-scores. It therefore reflects how consistently the brain is holding this tendency, not just “how big” the signal or metric is. A moderate gain held consistently over the session leads to a larger end-result in terms of the average amplitude or connectivity, hence the overall z-score.
There is a tendency in EEG biofeedback to think in terms of making “small things big” or making “big things small”. In fact, neurofeedback works primarily on a mechanism that is time-based, not size-based. By achieving and sustaining desired states more consistently, as a matter of habit and learning, the brain’s dynamics are changed, leading to improved self-regulation. In all, neurofeedback itself is very much a cerebral reflection of a game of golf, in which each “trial” of learning is an opportunity to re-orient and re-regulate the brain toward the desired goals.
